LTspice and Roman Blacks 2 transistor SMPS

[lk]

Greetings,

I find thishttp://www.romanblack.com/smps/smps.htm SMPS fascinating, so i made as close a copy as i could do in LT Spice, and i found something curious, the total power going into the circuit is around 6.5 watts and the energy output is 5,67 watts, that's around 87% efficiency which is think is very nice, but my problem is that Q3 in my Simulation(Q1 in RB's) is listed as having a dissipation of 6.1 watt, i don't understand this, it would surely destroy a to-92 package. For one if i cannot count on the energy dissipated in that transistor, it will be hard to make changes to the circuit, and then trust the numbers.


-lk

[c4757p]

I built the circuit in LTSpice with a 4.7V Zener (1N750) - 3.75V output - and a 100 Ohm load. I am getting a peak power in that transistor of 1.8W, but if I ctrl-click the waveform expression, it shows an average power of 6.33mW. The peak collector current is -430mA at startup and -220mA once it has stabilized, which is well within the maximum for most higher-power TO-92 transistors. The maximum power is related to the maximum temperature, so it's the average that will damage the transistor, not the peak.

[lk]

Thanks c4757p, i have been ctrl clicking my waveforms, and i can see that the collector current is averaging at 250mA(the bc327 is rated at 500 average) so i guess it should be alright even at a load of 500mA, but I'm still confused in regards to the shown power dissipation.

-lk

[HackedFridgeMagnet]

The average power of Q3 says 437mW from your simulation
So you are near the limit with a to92.

I have used this circuit to step down and power my controller ic on some boards, it works well and is very cheap but it takes a bit of soldering. I am only using about 70mA I think.

I seem to remember that, maybe?? Hero999 did some variation of it, and he is sometimes on this forum.

[c4757p]

I didn't noticed you include the LTspice file for the version you tried. I just downloaded it and played with it.

Try playing around with the base resistor for that transistor, it is not the best value. I got the power down to 500mW average with no detriment to the output quality by increasing it to 2k, and you could probably do better (my laptop is running on fumes right now, I'm out of time...)